AI isn't optimizing the funnel. It's dissolving it.

For thirty years, go-to-market has been organized around a funnel: a wide top of loosely-interested people, narrowing through stages, to a thin bottom of buyers. Every tool, metric, and team structure we built assumes that shape.

The funnel was never a law of nature. It was a consequence of two constraints: attention was scarce, and information was asymmetric. Buyers couldn’t easily find what they needed, so vendors built machinery to push them stage by stage.

AI removes both constraints at once.

What actually changed

A buyer with a capable model can now do, in an afternoon, the research that used to take a sales cycle: compare options, read between the lines of a pricing page, simulate an implementation, surface the objections your own SEs would have raised. The information asymmetry that justified half the funnel is gone.

What’s left isn’t a narrower funnel. It’s a different shape entirely — closer to a network of moments than a linear path. People arrive already late-stage on one dimension and cold on another. They self-serve the parts they can and want a human exactly where judgment is required.

The mistake teams are making

Most teams are pointing AI at the funnel: scoring leads faster, writing more sequences, summarizing more calls. That’s optimizing the shape instead of questioning it. You get a marginally more efficient version of a motion that’s quietly losing relevance.

The more interesting question: if buyers can now answer most of their own questions, what is the GTM org for? My current answer — and it’s a work in progress — is that the job shifts from manufacturing demand to removing friction from intent that already exists, and from gating information to being the most useful source of it.

That reframes everything downstream: what you measure, who you hire, what “content” even means. More on each of those soon.